Boosting COVID-19 Severity Detection with Infection-aware Contrastive Mixup Classifcation
share
This paper presents our solution for the 2nd COVID-19 Severity Detection Competition. This task aims to distinguish the Mild, Moderate, Severe, and Critical grades in COVID-19 chest CT images. In our approach, we devise a novel infection-aware 3D Contrastive Mixup Classifcation network for severity grading. Specifcally, we train two segmentation networks to frst extract the lung region and then the inner lesion region. The lesion segmentation mask serves as complementary information for the original CT slices. To relieve the issue of imbalanced data distribution, we further improve the advanced Contrastive Mixup Classifcation network by weighted cross-entropy loss. On the COVID-19 severity detection leaderboard, our approach won the frst place with a Macro F1 Score of 51.76 over 11.46
READ FULL TEXT
